In 2019, approximately 5 million individuals were diagnosed with gastrointestinal tract cancer globally, with about half eligible for radiation therapy. This treatment, crucial for many patients, faces challenges due to the manual segmentation process required in newer technologies like MR-Linacs. This paper propose a Se-ResNet50 Based Unet model for stomach and tract segmentation to ease the radiation therapy. The model is used architecture of UNet and a backbone network of Se-ResNet50. We train the model using the dataset from MR scan images from UW-Madison Carbone. The experiment has shown a promising result, our model achieves a Dice Coefficient score of 0.848, which outperforms other models like ResNet50, EfficientNetB0 and EfficentNetB1.